Figure 2 | Scientific Reports

Figure 2

From: Opportunistic assessment of ischemic heart disease risk using abdominopelvic computed tomography and medical record data: a multimodal explainable artificial intelligence approach

Figure 2

Performance of proposed Imaging + Clinical Fusion model compared to Pooled Cohort Equations (PCE), visualized through the receiver operating characteristic (ROC) curves (top row) and precision recall curves (bottom row) for (a) 1-year and (b) 5-year IHD risk modeling. Shaded lines indicate 95% confidence intervals (CI). Dashed lines show performance of a random (ROC curve) or a prevalence-based classifier (precision recall curve) as the simplest baselines. Area under the curve (AUC), 95% CI were determined using DeLong’s method for the ROC curve and using the bootstrap method for the precision-recall curve.

Back to article page